JP6244731B2 - Information processing apparatus and information processing program - Google Patents

Description

  The present invention relates to an information processing apparatus and an information processing program.

  In Patent Document 1, it is an object to enable language analysis with higher accuracy, and when analyzing a predetermined context, it is included in the context as local context information and the nearest word to be processed Further, the most recent word is a processing target, and as global context information, a plurality of words positioned before the processing target are processing targets, and the global context information is related to a context composed of a plurality of words in advance. A probability distribution is generated, a cluster ID is assigned, and it is determined to which probability distribution the global context information to be processed corresponds, the cluster ID of the probability distribution determined to be applicable, and the local distribution It is disclosed that context information is used, analysis of what kind of sentence is inputted, and it can be applied to an acoustic recognition device that performs language analysis.

  Patent document 2 aims at performing sentence speech recognition with high recognition accuracy and efficiency, and in the process of searching for a phrase sequence in the lattice, performs a semantic analysis of the phrase sequence obtained from the intermediate results of the analysis. Predict the meaning of what should be recognized from this semantic information and situation transition, and use this information as a constraint condition to eliminate phrase sequences with unexpected meanings in certain situations, It is disclosed that sentence analysis means for preventing the analysis from continuing due to the dropping of a phrase in the middle is incorporated in the recognition system.

  Patent Document 3 provides a speech recognition device, a speech recognition program, and a language model generation method and a language model generation device used in the speech recognition device, in which candidates for speech recognition are optimized and the speech recognition rate is improved. The speech recognition apparatus includes a language model storage unit storing a language model generated from a fixed pattern of speech, an acoustic model storage unit storing an acoustic model including acoustic characteristics of speech, a language model, and an acoustic model. And a speech processing unit that acoustically analyzes a speech signal and converts it into character information, and the language model is a speech model including character information acquired from a site specified by the URL of the organization to which the speaker belongs. It is generated from a fixed pattern, and the acoustic model is a model learned by telephone speech, and adopts such a language model and acoustic model. Because there can be converted to character information from the correct candidate, the speech recognition rate has been disclosed to be improved.

JP 2008-181537 A Japanese Unexamined Patent Publication No. 01-260494 JP 2005-208483 A

  An object of the present invention is to provide an information processing apparatus and an information processing program that use a document related to a speaker who utters speech of speech information and correct the recognition result of the speech information.

The gist of the present invention for achieving the object lies in the inventions of the following items.
The invention according to claim 1 is a first accepting means for accepting voice information, a second accepting means for accepting speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information, and the first Based on the recognition means for recognizing the speech information received by the receiving means, the acquisition means for acquiring the document related to the speaker identification information received by the second receiving means, and the document acquired by the acquisition means Correcting means for correcting the recognition result by the recognition means , wherein the correction means calculates a probability that a plurality of recognition results by the recognition means appear in the document acquired by the acquisition means, and the probability The information processing apparatus is characterized in that the recognition result by the recognition means is corrected based on the above .

The invention according to claim 2 is a first accepting means for accepting voice information, a second accepting means for accepting speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information, and the first Based on the recognition means for recognizing the speech information received by the receiving means, the acquisition means for acquiring the document related to the speaker identification information received by the second receiving means, and the document acquired by the acquisition means And a correcting means for correcting a recognition result by the recognizing means, wherein the recognizing means outputs a plurality of recognition results and a certainty factor for the recognition result for one phrase, When there are a plurality of recognition results with the certainty level higher or higher than a predetermined value for one clause, each of the plurality of recognition results and the recognition result of the clause before or after the clause, Combination Sega, calculates the probability of occurrence in the document obtained by the obtaining means, based on said probability is information processing apparatus characterized by modifying the recognition result by said recognizing means.

According to a third aspect of the present invention, the acquisition means is a document created by a speaker of the speaker identification information, and the document is created within a predetermined period from when the voice of the voice information is spoken. an information processing apparatus according to claim 1 or 2, characterized in that to obtain.

The invention of claim 4 is a first receiving means for receiving voice information, a second receiving means for receiving speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information, Recognizing means for recognizing voice information received by the first receiving means, acquiring means for acquiring a document related to the speaker identification information received by the second receiving means, and a document acquired by the acquiring means Based on the recognition means functioning as a correction means for correcting the recognition result by the recognition means, the correction means calculates a probability that a plurality of recognition results by the recognition means appear in the document acquired by the acquisition means. An information processing program for correcting a recognition result by the recognition means based on the probability .
According to a fifth aspect of the present invention, the computer includes a first receiving unit that receives voice information, a second receiving unit that receives speaker identification information that can uniquely identify a speaker who has spoken the voice of the voice information, Recognizing means for recognizing voice information received by the first receiving means, acquiring means for acquiring a document related to the speaker identification information received by the second receiving means, and a document acquired by the acquiring means Based on the recognition means functioning as a correction means for correcting the recognition result by the recognition means, the recognition means outputs a plurality of recognition results and a certainty factor for the recognition result for one phrase, and the correction means When there are a plurality of recognition results with the certainty level higher or higher than a predetermined value for one phrase, each of the plurality of recognition results and a sentence before or after the phrase An information processing program that calculates a probability that a combination with a recognition result of the above appears in a document acquired by the acquisition unit, and corrects the recognition result by the recognition unit based on the probability .

According to the information processing apparatus of the first aspect, the recognition result of the voice information can be corrected using the document related to the speaker who uttered the voice of the voice information. In addition, the recognition result can be corrected using the probability that a plurality of recognition results appear in the document.

According to the information processing apparatus of the second aspect , the recognition result of the voice information can be corrected using the document related to the speaker who uttered the voice of the voice information. In addition, the recognition result can be corrected more efficiently than when the configuration is not provided.

According to the information processing apparatus of the third aspect , the recognition result can be corrected using a document created by a recent speaker.

According to the information processing program of claim 4 , the recognition result of the voice information can be corrected using the document related to the speaker who uttered the voice of the voice information. In addition, the recognition result can be corrected using the probability that a plurality of recognition results appear in the document.
According to the information processing program of the fifth aspect, the recognition result of the voice information can be corrected using the document related to the speaker who uttered the voice of the voice information. In addition, the recognition result can be corrected more efficiently than when the configuration is not provided.

It is a conceptual module block diagram about the structural example of this Embodiment. It is a flowchart which shows the process example by this Embodiment. It is explanatory drawing which shows the data structure example of a document table. It is explanatory drawing which shows the example of a data structure of a meeting table. It is a flowchart which shows the process example by this Embodiment. It is a flowchart which shows the process example by this Embodiment. It is a block diagram which shows the hardware structural example of the computer which implement | achieves this Embodiment.

Hereinafter, an example of a preferred embodiment for realizing the present invention will be described with reference to the drawings.
FIG. 1 shows a conceptual module configuration diagram of a configuration example of the present embodiment.
The module generally refers to components such as software (computer program) and hardware that can be logically separated. Therefore, the module in the present embodiment indicates not only a module in a computer program but also a module in a hardware configuration. Therefore, the present embodiment is a computer program for causing these modules to function (a program for causing a computer to execute each procedure, a program for causing a computer to function as each means, and a function for each computer. This also serves as an explanation of the program and system and method for realizing the above. However, for the sake of explanation, the words “store”, “store”, and equivalents thereof are used. However, when the embodiment is a computer program, these words are stored in a storage device or stored in memory. It is the control to be stored in the device. Modules may correspond to functions one-to-one, but in mounting, one module may be configured by one program, or a plurality of modules may be configured by one program, and conversely, one module May be composed of a plurality of programs. The plurality of modules may be executed by one computer, or one module may be executed by a plurality of computers in a distributed or parallel environment. Note that one module may include other modules. Hereinafter, “connection” is used not only for physical connection but also for logical connection (data exchange, instruction, reference relationship between data, etc.). “Predetermined” means that the process is determined before the target process, and not only before the process according to this embodiment starts but also after the process according to this embodiment starts. In addition, if it is before the target processing, it is used in accordance with the situation / state at that time or with the intention to be decided according to the situation / state up to that point. When there are a plurality of “predetermined values”, the values may be different from each other, or two or more values (of course, including all values) may be the same. In addition, the description having the meaning of “do B when it is A” is used in the meaning of “determine whether or not it is A and do B when it is judged as A”. However, the case where it is not necessary to determine whether or not A is excluded.
In addition, the system or device is configured by connecting a plurality of computers, hardware, devices, and the like by communication means such as a network (including one-to-one correspondence communication connection), etc., and one computer, hardware, device. The case where it implement | achieves by etc. is also included. “Apparatus” and “system” are used as synonymous terms. Of course, the “system” does not include a social “mechanism” (social system) that is an artificial arrangement.
In addition, when performing a plurality of processes in each module or in each module, the target information is read from the storage device for each process, and the processing result is written to the storage device after performing the processing. is there. Therefore, description of reading from the storage device before processing and writing to the storage device after processing may be omitted. Here, the storage device may include a hard disk, a RAM (Random Access Memory), an external storage medium, a storage device via a communication line, a register in a CPU (Central Processing Unit), and the like.

  The information processing apparatus 100 according to the present embodiment recognizes voice, and as shown in the example of FIG. 1, the voice reception module 110, the voice recognition module 120, the participant acquisition module 130, and the document storage module 140. , A document acquisition module 150, a speech recognition result correction module 160, and a speech recognition result output module 170. The information processing apparatus 100 is used in a meeting such as a meeting where a plurality of people gather and recognizes the speech (voice information) of the attendee at the meeting.

  The voice reception module 110 is connected to the voice recognition module 120. The voice reception module 110 receives voice information. For example, the voice information of the speaker is received from a microphone. Specifically, the microphone may be built in a mobile terminal (for example, a mobile phone including a smartphone). The audio information may be already recorded. That is, it includes reading out audio information stored in a hard disk (including those connected via a network in addition to those built in the information processing apparatus 100).

  The voice recognition module 120 is connected to the voice reception module 110 and the voice recognition result correction module 160. The voice recognition module 120 recognizes voice information received by the voice reception module 110. Conventional speech recognition technology may be used. For example, the speech recognition result may be output in a phrase lattice format. Further, the speech recognition result may be output in a format other than the phrase lattice format, for example, a confusion network. The phrase lattice format and the confusion network include a plurality of recognition results for one phrase and certainty about the recognition result. That is, the speech recognition module 120 may output a plurality of recognition results and a certainty factor for the recognition results for one phrase. The certainty factor expresses the degree of correctness of the recognition result by, for example, a numerical value ranging from 0 to 1. For example, when the voice information received by the voice reception module 110 is compared with a pattern that is a dictionary, and there is no difference (matches in the feature space), the certainty factor is high.

  The participant acquisition module 130 is connected to the document acquisition module 150. The participant acquisition module 130 receives speaker identification information (user ID (IDentification)) that can uniquely identify a speaker who has spoken the voice of the voice information received by the voice reception module 110 in the present embodiment. For example, immediately before speaking, during speaking or immediately after speaking, attendees of the meeting may input their user IDs using a keyboard or the like, or an IC card or portable terminal owned by the attendees The user ID may be read from the above. When using the microphone built in the portable terminal, the user ID built in the portable terminal may be extracted with the utterance as an opportunity. Alternatively, the voice print of the attendee may be acquired in advance, and the user ID that matches the voice print of the voice information may be extracted. Further, the face of the participant who is speaking may be recognized from an image taken by a camera installed in the conference room, and the user ID may be acquired. Furthermore, an attendee to the meeting may be identified from an electronic conference room reservation system or the like, and the face of the attendee may be recognized from within the identified attendee, and the user ID may be acquired.

The document storage module 140 is connected to the document acquisition module 150. The document storage module 140 stores a document. Information for managing the document is also stored. For example, the document table 300 is stored. FIG. 3 is an explanatory diagram showing an example of the data structure of the document table 300. The document table 300 includes a document ID column 310, a document name column 320, a creator ID column 330, a creation date / time column 340, an editor ID column 350, an edit date / time column 360, a viewer ID column 370, and a browse date / time column 380. ing. The document ID column 310 stores information (document ID) for uniquely identifying a document in the present embodiment. The document name column 320 stores the document name of the document. The creator ID column 330 stores information (creator ID, user ID) for uniquely identifying the creator (user) of the document in the present embodiment. The creation date / time column 340 stores the creation date / time (year, month, day, hour, minute, second, second or less, or a combination thereof) when the document was created. The editor ID column 350 stores information (editor ID, user ID) for uniquely identifying the editor (user) of the document in the present embodiment. The edit date / time column 360 stores the edit date / time when the document was edited. The viewer ID column 370 stores information (viewer ID, user ID) for uniquely identifying the viewer (user) of the document in the present embodiment. The browsing date / time column 380 stores the browsing date / time when the document was browsed.
When the document is edited a plurality of times, there are as many sets of the editor ID column 350 and the edit date / time column 360 as the number of times. Similarly, when the document is browsed a plurality of times, there are a set of the viewer ID column 370 and the browse date / time column 380 corresponding to the number of times.
The document stored in the document storage module 140 may include a document related to a speaker or a conference participant (specifically, created, edited, viewed, etc.). For example, there are presentation materials and minutes.

The document acquisition module 150 is connected to the participant acquisition module 130, the document storage module 140, and the speech recognition result correction module 160. The document acquisition module 150 acquires a document related to the speaker identification information received by the participant acquisition module 130 from the document storage module 140. Here, the document related to the speaker identification information includes any one or a combination of a document created, edited, and viewed by the speaker of the speaker identification information. Moreover, you may give priority. For example, the speaker may add the weighting coefficient for correcting the recognition result by the speech recognition result correction module 160 in the order of the created document, the edited document, and the viewed document.
Further, the document acquisition module 150 may acquire a document created by a speaker of the speaker identification information and created within a predetermined period from when the voice information is spoken. Good. “When uttered” means that when the voice reception module 110 receives voice information, and the date and time of utterance is attached to the voice information, the date and time may be extracted.
Further, the document acquisition module 150 may acquire a document related to the conference participant identification information (participant ID) from the document storage module 140. Of course, the speaker identification information is included in the participant identification information. For example, participant identification information is acquired from the conference table 400. FIG. 4 is an explanatory diagram showing an example of the data structure of the conference table 400. The conference table 400 includes a conference ID column 410, a conference name column 420, a date / time column 430, and a participant ID column 440. The conference ID column 410 stores information (conference ID) for uniquely identifying a conference in the present embodiment. The meeting name column 420 stores the meeting name of the meeting. The date and time column 430 stores the date and time when the meeting was held. The participant ID column 440 stores a participant ID (user ID) who participated in the conference. The document acquisition module 150 acquires from the conference table 400 the participant ID in the conference in which the audio information received by the information processing apparatus 100 is spoken. Whether or not the participant ID matches the creator ID, editor ID, and viewer ID stored in the creator ID column 330, the editor ID column 350, and the viewer ID column 370 of the document table 300. And the document is acquired from the matched document ID. The conference table 400 may be stored in the information processing apparatus 100, or may be read via a communication line from what is stored in a conference management apparatus that performs conference room reservation management or the like. The above-described weighting process may be applied to a document related to the conference participant ID.

The voice recognition result correction module 160 is connected to the voice recognition module 120, the document acquisition module 150, and the voice recognition result output module 170. The voice recognition result correction module 160 corrects the recognition result by the voice recognition module 120 based on the document acquired by the document acquisition module 150.
Further, the speech recognition result correction module 160 calculates the probability that a plurality of recognition results by the speech recognition module 120 will appear in the document acquired by the document acquisition module 150, and based on the probability, the speech recognition module 120 The recognition result may be corrected.
In addition, when there are a plurality of recognition results whose certainty factor is higher than or higher than a predetermined value for one phrase, the speech recognition result correction module 160 includes each of the plurality of recognition results and the phrase. The probability that the combination with the phrase recognition result before or after the phrase appears in the document acquired by the document acquisition module 150 is calculated, and the recognition result by the speech recognition module 120 is corrected based on the probability. May be.
For example, there are a clause A, a clause B, and a clause C. For the clause A and the clause C, there is one with a certainty factor higher than a predetermined value, and the recognition results are A1 and C1, respectively. When there are a plurality of recognition results B1 and B2 for the phrase B and both are higher than a predetermined value,
(1) A1 B1 C1
(2) A1 B2 C1
Create a combination of Then, the probability that the combination (1) occurs in the document acquired by the document acquisition module 150 and the probability that the combination (2) occurs in the document are obtained.
Further, as described above, the speech recognition result correction module 160 may calculate the probability using the weight for each document. For example, the final probability may be obtained by multiplying the probability by a weight.

  The voice recognition result output module 170 is connected to the voice recognition result correction module 160. The voice recognition result output module 170 outputs the voice recognition result corrected by the voice recognition result correction module 160. Outputting the speech recognition result includes, for example, displaying on a display device such as a display, writing the speech recognition result in a storage device such as a conference database, storing it in a storage medium such as a memory card, and a minutes creation device And the like, and the like.

FIG. 2 is a flowchart showing an example of processing according to this embodiment.
In step S202, the voice reception module 110 receives voice data with a speaker ID.
In step S204, the voice recognition module 120 performs voice recognition of the voice data received in step S202.
In step S206, the participant acquisition module 130 acquires a conference participant ID.
In step S <b> 208, the document acquisition module 150 acquires a document in which the participant with the participant ID is involved from the document storage module 140. The participant ID includes a speaker ID. Therefore, a document in which at least the speaker with the speaker ID is involved is acquired.
In step S210, the speech recognition result correction module 160 corrects the speech recognition result using the document acquired in step S208. Details will be described later with reference to the flowchart shown in the example of FIG.
In step S212, the speech recognition result output module 170 outputs the speech recognition result corrected in step S210.
Note that any of the processing in step S202 and step S204 and the processing in step S206 and step S208 may be performed first or may be performed in parallel.

FIG. 5 is a flowchart showing an example of processing according to the present embodiment. A detailed processing example (1) of step S210 will be described.
In step S502, one recognition result is extracted from a plurality of recognition results for the phrase. For example, the following is accepted as a voice recognition result. Each row shows a plurality of recognition results for one clause (speech information of one section). Here, “-” represents a blank.
・ (Confirmation) (Each) (Evil)
・ (−) (2)
(No) (-) (na) (ga)
・ (Work) (Industry)
・ (To)
・ (It depends) (It takes)
・ (Course) (Man-hours) (Hose)
・ (Ga) (ka) (house) (family)
One recognition result is extracted from each recognition result. For example, (confirmation) is extracted from “(confirmation) (each) (bad person)”. One by one is extracted from other lines (a plurality of recognition results for the same phrase).

In step S504, a set of recognition results extracted in step S502 is generated. In the above-described example, a pair of “(confirmation) (−) (of) (work) (to) (related) (course) (ga)” is generated. Of course, the process returns in step S508, and the second and subsequent processes are different combinations. For example, “(each) (−) (no) (work) (ni) (related) (course) (ga)”.
In step S506, the appearance probability of the generated character string set in the document is calculated. The target document here is the document acquired in step S208.
Appearance probability is calculated as follows.
Appearance probability = number of documents in which all words appear / number of documents

In step S508, it is determined whether all combinations have been generated. If they have been generated, the process proceeds to step S510, and otherwise, the process returns to step S502. That is, the processing from step S502 to step S506 is performed on all combinations extracted one by one from among a plurality of recognition results for the same section.
In step S510, the recognition result is confirmed. Specifically, a combination of recognition results having the highest appearance probability is determined as the recognition result.
In step S512, the confirmed recognition result is transferred to the speech recognition result output module 170.

FIG. 6 is a flowchart showing an example of processing according to this embodiment. A detailed processing example (2) of step S210 will be described. In the process of the flowchart shown in the example of FIG. 5, the certainty factor of the recognition result is not used, but in FIG. 6, a processing example using the certainty factor will be described.
In step S602, a recognition result having a certainty factor equal to or higher than a threshold (predetermined value) is extracted.
As an example of the phrase lattice shown below, “confirmation”, “case”, and “bad person” are output as speech recognition results, and the certainty levels of the phrases are 0.922 and 0.042, respectively. , 0.037 is output.
Similarly, “−” and “2” are output in the second clause. Here, “-” represents a blank.
・ (Confirmation: 0.922) (Each: 0.042) (Bad: 0.037)
・ (-: 0.958) (2: 0.042)
・ (No: 0.823) (-: 0.094) (na: 0.040) (ga: 0.034)
・ (Work: 0.952) (Industry: 0.048)
・ (To: 1.000)
・ (According: 0.592) (According: 0.409)
・ (Course: 0.825) (Man-hour: 0.120) (Hose: 0.055)
・ (Ga: 0.899) (or: 0.051) (house: 0.026) (family: 0.024)
From the phrase lattice that is the result of speech recognition, a phrase having a certainty of the phrase equal to or greater than a threshold is selected. The following is an example of selection when the threshold value is “0.1”.
・ (Confirmation: 0.922)
・ (-: 0.958)
・ (No: 0.823)
・ (Work: 0.952)
・ (To: 1.000)
・ (According: 0.592) (According: 0.409)
・ (Course: 0.825) (Man-hour: 0.120)
・ (G: 0.899)

  In step S604, it is determined whether or not there are a plurality of extracted recognition results for one phrase. If there are a plurality of recognition results, the process proceeds to step S606, otherwise the process proceeds to step S612. In other words, if there is one phrase with a certainty level greater than or equal to the threshold, that is used as the final analysis result of the phrase. In the above example, (confirmation: 0.922), (-: 0.958), (no: 0.823), (operation: 0.952), (ni: 1.000), (but: 0.0. 899) is that. If there are a plurality of phrases having a certainty level equal to or greater than the threshold value in the same phrase, the processes after step S606 are performed.

In step S606, each of the confirmed recognition result (one phrase with a certainty level greater than or equal to a threshold value) and an indeterminate recognition result (a plurality of phrases with a certainty level greater than or equal to a threshold value) is configured. Create a tuple. That is, the probability that the word whose analysis result is confirmed and the candidate word appears in the document acquired in step S208 is calculated. At this time, all confirmed words may be used. Moreover, you may limit only to an independent word, a noun, and an unknown word using a part of speech information. That is, particles such as “ni”, auxiliary verbs, and the like may be excluded from the target.
To generate a set, the previous or subsequent recognition result is combined with the indeterminate recognition result.
In the above-described example, for each of “related” and “related”, a pair of “confirmation, work, related” and “confirmation, related work” is generated. Here, at the position where the uncertain recognition result appears, a pair with the previously recognized recognition result is generated. Then, after confirming the indeterminate recognition result in step S610, the process may proceed to the next indeterminate recognition result, generate a set, and repeat this until the final recognition result.
Also, here, all combinations, “confirmation, work, ga, course, ga”, “confirmation, work, gait, course, ga”, “confirmation, It may be made to generate “work, work, man-hours”, “confirmation, work, work, man-hours”.

In step S608, the appearance probability of the generated character string set in the document is calculated.
In the above-described example, the probability of occurrence of “confirmation of work, related to” and “confirmation of work related to” is calculated. The formula for calculating the appearance probability is as described above.
For example, here, it is assumed that the occurrence probability of “confirmation, work, related” is 0.004, and the appearance probability of “confirmation, work, related” is 0.012.
In step S610, the recognition result is confirmed.
In the above example, the occurrence probability of “takes” (“confirmation, work, take”) is higher than the occurrence probability of “related” (“confirmation, work, take”). “Such” is adopted as the final analysis result. At this time, the one with the maximum appearance probability may be adopted, or the one with the appearance probability higher than the threshold may be adopted. In the latter case, the appearance probability is calculated again after the ambiguity of other phrases is first resolved. Similarly, the final output result is determined for the remaining clauses. When the ambiguity is resolved for all the clauses, the process is terminated.
In step S612, the recognition result determined in step S610 is passed to the speech recognition result output module 170.

  The hardware configuration of the computer on which the program according to the present embodiment is executed is a general computer, specifically a computer that can be a personal computer or a server, as illustrated in FIG. That is, as a specific example, the CPU 701 is used as a processing unit (calculation unit), and the RAM 702, ROM 703, and HD 704 are used as storage devices. For example, a hard disk may be used as the HD 704. A CPU 701 that executes programs such as the voice reception module 110, the voice recognition module 120, the participant acquisition module 130, the document acquisition module 150, the voice recognition result correction module 160, the voice recognition result output module 170, and the like, and stores the programs and data. A RAM 702, a ROM 703 storing a program for starting up the computer, an HD 704 as an auxiliary storage device (may be a flash memory or the like) having a function as the document storage module 140, a keyboard, a mouse A communication device for connecting to a receiving device 706 that accepts data based on a user's operation on a touch panel, an output device 705 such as a CRT or a liquid crystal display, and a communication network such as a network interface card. Interface 707, and, and a bus 708 for exchanging data by connecting them. A plurality of these computers may be connected to each other via a network.

Among the above-described embodiments, the computer program is a computer program that reads the computer program, which is software, in the hardware configuration system, and the software and hardware resources cooperate with each other. Is realized.
Note that the hardware configuration illustrated in FIG. 7 illustrates one configuration example, and the present embodiment is not limited to the configuration illustrated in FIG. 7, and is a configuration capable of executing the modules described in the present embodiment. I just need it. For example, some modules may be configured by dedicated hardware (for example, ASIC), and some modules may be in an external system and connected via a communication line. A plurality of systems shown in FIG. 5 may be connected to each other via communication lines so as to cooperate with each other. In particular, in addition to personal computers, information appliances, copiers, fax machines, scanners, printers, and multifunction machines (image processing apparatuses having two or more functions of scanners, printers, copiers, fax machines, etc.) Etc. may be incorporated.

  Further, in the description of the above-described embodiment, “more than”, “less than”, “greater than”, and “less than (less than)” in a comparison with a predetermined value contradicts the combination. As long as the above does not occur, “larger”, “smaller (less than)”, “more than”, and “less than” may be used.

The program described above may be provided by being stored in a recording medium, or the program may be provided by communication means. In that case, for example, the above-described program may be regarded as an invention of a “computer-readable recording medium recording the program”.
The “computer-readable recording medium on which a program is recorded” refers to a computer-readable recording medium on which a program is recorded, which is used for program installation, execution, program distribution, and the like.
The recording medium is, for example, a digital versatile disc (DVD), which is a standard established by the DVD Forum, such as “DVD-R, DVD-RW, DVD-RAM,” and DVD + RW. Standard “DVD + R, DVD + RW, etc.”, compact disc (CD), read-only memory (CD-ROM), CD recordable (CD-R), CD rewritable (CD-RW), Blu-ray disc ( Blu-ray (registered trademark) Disc), magneto-optical disk (MO), flexible disk (FD), magnetic tape, hard disk, read-only memory (ROM), electrically erasable and rewritable read-only memory (EEPROM (registered trademark)) )), Flash memory, Random access memory (RAM) SD (Secure Digital) memory card and the like.
The program or a part of the program may be recorded on the recording medium for storage or distribution. Also, by communication, for example, a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a wired network used for the Internet, an intranet, an extranet, etc., or wireless communication It may be transmitted using a transmission medium such as a network or a combination of these, or may be carried on a carrier wave.
Furthermore, the program may be a part of another program, or may be recorded on a recording medium together with a separate program. Moreover, it may be divided and recorded on a plurality of recording media. Further, it may be recorded in any manner as long as it can be restored, such as compression or encryption.

DESCRIPTION OF SYMBOLS 100 ... Information processing apparatus 110 ... Voice reception module 120 ... Voice recognition module 130 ... Participant acquisition module 140 ... Document storage module 150 ... Document acquisition module 160 ... Voice recognition result correction module 170 ... Voice recognition result output module

Claims (5)

First receiving means for receiving voice information;
Second receiving means for receiving speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information;
Recognizing means for recognizing voice information received by the first receiving means;
Obtaining means for obtaining a document related to the speaker identification information received by the second receiving means;
Correction means for correcting the recognition result by the recognition means based on the document acquired by the acquisition means ;
The correction means calculates a probability that a plurality of recognition results by the recognition means appear in the document acquired by the acquisition means, and corrects the recognition results by the recognition means based on the probabilities. Information processing apparatus. First receiving means for receiving voice information;
Second receiving means for receiving speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information;
Recognizing means for recognizing voice information received by the first receiving means;
Obtaining means for obtaining a document related to the speaker identification information received by the second receiving means;
Correction means for correcting the recognition result by the recognition means based on the document acquired by the acquisition means ;
The recognition means outputs a plurality of recognition results for one phrase and a certainty factor for the recognition results,
When there are a plurality of recognition results for which the certainty factor is higher or higher than a predetermined value for one clause, each of the plurality of recognition results and before or after the clause. An information processing apparatus that calculates a probability that a combination with a recognition result of a phrase appears in a document acquired by the acquisition unit, and corrects the recognition result by the recognition unit based on the probability . The acquisition means acquires a document created by a speaker of the speaker identification information, the document created within a predetermined period from when the voice of the voice information is spoken. The information processing apparatus according to claim 1 or 2 . Computer
First receiving means for receiving voice information;
Second receiving means for receiving speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information;
Recognizing means for recognizing voice information received by the first receiving means;
Obtaining means for obtaining a document related to the speaker identification information received by the second receiving means;
Based on the document acquired by the acquisition means, function as a correction means for correcting the recognition result by the recognition means ,
The correction means calculates a probability that a plurality of recognition results by the recognition means appear in the document acquired by the acquisition means, and corrects the recognition results by the recognition means based on the probabilities.
An information processing program characterized by that . Computer
First receiving means for receiving voice information;
Second receiving means for receiving speaker identification information capable of uniquely identifying a speaker who has spoken the voice of the voice information;
Recognizing means for recognizing voice information received by the first receiving means;
Obtaining means for obtaining a document related to the speaker identification information received by the second receiving means;
Based on the document acquired by the acquisition means, function as a correction means for correcting the recognition result by the recognition means ,
The recognition means outputs a plurality of recognition results for one phrase and a certainty factor for the recognition results,
When there are a plurality of recognition results for which the certainty factor is higher or higher than a predetermined value for one clause, each of the plurality of recognition results and before or after the clause. The probability that the combination with the recognition result of the phrase appears in the document acquired by the acquisition unit is calculated, and the recognition result by the recognition unit is corrected based on the probability.
An information processing program characterized by that .

Images